The Extent and Impact of Food Non-Tariff
Barriers in Rich Countries
by
Scott Bradford
Paper presented at the
International Agricultural Trade Research Consortium
Annual Meeting Theme Day
December 4-6, 2005
San Diego, California
THE EXTENT AND IMPACT OF FOOD NON-TARIFF
BARRIERS IN RICH COUNTRIES
Scott Bradford
Brigham Young University
January 2006
International trade negotiations have significantly reduced food tariffs in rich
countries, increasing the relative importance of non-tariff barriers (NTBs). Since
reducing them often requires deeper integration, the resulting negotiations have been
more fractious and difficult than earlier efforts. The Uruguay Round took almost eight
years, by far the longest round on record. The Doha Round has faced trouble in
Seattle, Cancun, and Hong Kong. Given these considerations, we need to weigh the
benefits of reducing NTBs. If these benefits are small, then perhaps the time has come
to place a lower priority on achieving deeper economic integration. On the other hand,
if the barriers remain substantial, it could be worthwhile to invest considerable political
capital in their elimination.
This paper presents a new method for estimating tariff equivalents of NTBs for
final food goods in OECD countries. The analysis exploits detailed, comprehensive, and
careful price comparisons: matched retail prices that the OECD collects on a regular
basis in order to calculate purchasing power parity (PPP) estimates. Since this method
does not identify policies, I strive to supplement the numbers by presenting preliminary
information on possible sources of the barriers. I then use an applied general
equilibrium model to provide a broad-brushed assessment of the impact of these NTBs.
The results imply that NTBs significantly restrict trade in OECD nations and that
removing them would bring large gains to them and to developing countries. Thus, this
research implies that continued efforts to negotiate the reduction of NTBs will indeed
exceed the costs.
This work draws on joint work with Robert Lawrence, to whom I am greatly indebted. I thank David
Blandford, Michael Ferrantino, Tim Josling, Suchada Langley, Christine McDaniel, William Meyers, Pat
Westhoff, and Linda Young for their help. I also thank participants in the BYU Economics Department
Raw Research Seminar, the APEC Non-tariff Measure Workshop (Bangkok, October 2003), the Empirical
Trade Analysis Conference (Washington, DC, January 2004), the 2004 AAEA annual meeting in Denver,
and the 2005 IATRC annual meeting in San Diego. I take full responsibility for any errors in this paper.
Page 1 of 39
1
INTRODUCTION
International trade negotiations over the past few decades have greatly reduced
food 1 tariffs in rich nations, leading to a commensurate increase in the relative
importance of food non-tariff barriers (NTBs). This has presented two challenges for
trade analysts and negotiators alike. First, since NTBs cannot be measured as easily as
tariffs, we have become less sure about how much food protection rich nations have.
Second, since NTBs lack tariffs’ transparency and are often embedded within complex
domestic regulatory regimes, reducing these NTBs generally requires more work than
reducing tariffs does.
This extra work stems not just from more difficult and technical subject matter
but also from more intense political opposition to deeper integration. The Uruguay
Round took almost eight years, the longest round on record, in part because it reduced
NTBs more than any other previous round. The Doha Round, which seeks major
opening in food sectors, suffered setbacks in Seattle and Cancun and had to settle for
disappointing results in Hong Kong mainly because of a stalemate on agriculture. Food
producer interests remain politically strong in rich nations, even to the point of possibly
preventing a global trade deal that would bring great benefits to the world.
Despite this opposition, the desire for more integration still drives policy, and
support for significant liberalization in agriculture and food products is widespread.
Many nations continue to negotiate bilateral and regional agreements, which almost
always call for substantial barrier reductions in food. The United States has moved
1
By “food” I mean agricultural products; fishery products; and processed food, drink, and tobacco
products.
Page 2 of 39
beyond free trade agreements (FTAs) with Canada (CUSFTA) and Mexico (NAFTA) and
concluded FTAs with at least 10 other nations. Several other deals are in the works.
Even Japan has gotten into the FTA act lately, and there are signs there that it may
soon begin to open agriculture as part of multilateral and bilateral deals.
Given strong support for, and opposition to, reducing food NTBs, we need to
weigh the benefits of doing so. If they are small, then perhaps we should place a lower
priority on achieving deeper economic integration. On the other hand, large potential
gains could make it worthwhile to invest considerable political capital in the reduction of
food NTBs.
Assessing the worth of food NTB reduction involves two tasks: 1) Reliably
measuring the height of the NTBs, and 2) Using an economic model to infer the
potential economic gains from their removal. Accordingly, I first present a new method
for estimating tariff equivalents of NTBs for food final goods in OECD economies. The
analysis exploits detailed, comprehensive, and careful price comparisons. I also present
some preliminary information on the policies behind the estimates. Then, I use an
applied general equilibrium (AGE) model to provide a broad-brushed assessment of the
impact of these NTBs. 2 The results imply that food final goods NTBs greatly restrict
trade in many OECD nations, especially in Japan, and that removing them would bring
large gains to the world economy, for rich and poor countries alike. Thus, this research
2
This analysis gives an overview of the size and shape of the protection “forest”, without detailed
descriptions of individual “trees”. Assessing the effects of particular policies, however, is important future
work since it would probably facilitate the negotiations that this paper implies are worthwhile.
Page 3 of 39
implies that continued efforts to negotiate the reduction of food NTBs will indeed
exceed the costs.
2
MEASURING NTBs
Nations can protect their markets in many different ways, making it hard to
determine just how much protection different industries enjoy. As trade agreements
have brought tariff reductions, governments have relied on a variety of more opaque
but effective tools for insulating domestic food markets from foreign competition. For
the purposes of this paper, I will define NTBs as follows: “A government policy or
practice, other than a tariff, that raises the domestic price of a good above its import
price”. Note that this definition does not include subsidies, since they do not drive
wedges between domestic and import prices, even though such subsidies could restrict
imports. This definition encompasses barriers that drive wedges between prices.
Such NTBs include quotas and the procedures used to administer them; regulations
that limit or completely exclude imports, such as sanitary and phytosanitary restrictions,
testing and certification standards, labeling and packaging requirements, and food
additive rules; inadequate protection for trademarks and geographical indications;
restrictive distribution systems; burdensome customs procedures; safeguards, including
anti-dumping duties; biased government procurement; rules of origin; sanctions; and
threats of protection. Even when not created with protectionist intent, these policies
can inhibit international arbitrage, shield producers, and shrink the world economy.
Page 4 of 39
2.1 Other Approaches to Measuring NTBs
This section discusses three prominent NTB measurement approaches that have
been applied to food: 1) Counting NTBs and computing coverage ratios, 2) Inferring
protection from trade flows, and 3) Inferring protection from price gaps. Then, this
paper’s method is discussed.
2.1.1 Compute NTB “Coverage Ratios”
The United Nations has developed “NTB coverage ratios” by computing what
percentage of products within a sector has an NTB. Unfortunately, this measure does
not take account of how restrictive each barrier is. One sector may have many
products that are subject to minor NTBs. Another sector may have just a few products
with very restrictive NTBs. The first sector would have a much higher NTB coverage
ratio, while we would expect the second sector to actually have more restrictive trade
barriers. Also, the UN’s accounting probably does not cover all NTBs. For instance,
these coverage ratios do not include inefficient customs procedures, even though they
probably significantly restrict a wide variety of imports.
2.1.2 Infer Protection from Trade Flows
This approach seeks to measure the effects of NTBs by estimating their impact
on the volume of trade in different industries. Researchers use models to predict trade
patterns absent any barriers (on the basis of factors such as country size, distance from
other economies, and factor endowments) and then use the gap between actual and
Page 5 of 39
predicted trade flows to infer protection. This method has the advantage of being able
to capture the aggregate impact of all barriers combined, even ones not considered by
NTB list-makers. 3 This approach, however, depends on having a trade model that can
accurately account for all determinants of trade, besides barriers, which is an ambitious
requirement. One wonders how much of the gap between predicted and actual flows
results from barriers and how much results from model misspecification or data
mismeasurement or both. The fact that one has to specify demand elasticities in order
to convert the quantity shortfalls into tariff-equivalents introduces another source of
uncertainty.
2.1.3 Price Gaps
Like the second approach, this method has the virtue of capturing the full impact
of all NTBs. It has the additional virtues of not relying on any single model and
providing tariff-equivalent measures directly. Although it has pitfalls, I believe that the
price gap approach has the most promise for measuring NTBs. With many possible
barriers to trade, I believe that one can best account for all of them by using the
information that prices concisely convey.
The basic philosophy behind this approach is that barriers to arbitrage across
national borders should be considered barriers to trade. 4 If international markets are
integrated, sellers cannot raise domestic prices above prices that would attract
3
One popular version of this approach is to use so-called gravity equations. For an excellent review of
this methodology, see Frankel 1997.
4
This does not depend on individual consumers engaging in arbitrage. Organized and well-informed
trading companies and other international wholesalers can easily seize arbitrage opportunities.
Page 6 of 39
arbitrage from abroad. One needs to carefully account for unavoidable costs associated
with shipping goods between economies. Once one has done this, however, if a price
gap exists for equivalent goods in two different nations, then one can conclude that the
higher-priced market is protected. Moreover, one can use the price gap as a measure
of the extent of protection. Thus, a single number can give the total effect of all trade
barriers. These gaps may be caused in part by policies that are not explicitly designed
to impede trade, such as overly harsh sanitary standards. No matter what the intent,
however, which can be difficult to judge anyway, I presume that policies that segment
national markets are trade barriers. 5
The key to using this approach is obtaining appropriate price measures. Such
efforts confront three challenges. The first is comparing prices of equivalent goods.
Even if they have the same name, goods may have very different levels of quality.
Thus, surveyors need to work hard to ensure comparability. Many researchers have
used unit values as price proxies because they are widely available. These can provide
reasonable estimates of price gaps at very detailed classification levels (eg, Harmonized
System 10-digit), but, at higher levels of aggregation, unit values tend to be notoriously
inexact measures of prices because of large quality differences in products.
A second challenge is using producer, rather than consumer, prices. Most price
surveys are undertaken with a view to comparing costs to the consumer. In order to
accurately gauge protection for producers, though, one should compare producer
5
This notion corresponds to that of Knetter and Goldberg 1996, which argues that “A market is
segmented if the location [sic] of the buyers and the sellers influences the terms of the transaction in a
substantial way (i.e. by more than the marginal cost of physically moving the good from one location to
another).” (pp 3-4.)
Page 7 of 39
prices. Data gathered at the retail level include non-traded value added, such as
distribution margins and transportation costs. These prices may therefore provide an
inaccurate picture of protection since they include elements that cannot be eliminated
through arbitrage. The price of a pound of coffee purchased in a supermarket in Tokyo
may be higher than a pound of the same brand of coffee purchased in New York, either
because trade barriers raise the wholesale price of coffee or because the costs of
distributing coffee in Tokyo are higher, or both. One who seeks to isolate the role of
trade barriers needs to compare producer, rather than consumer, prices.
A third challenge relates to the comprehensiveness of coverage. Samples of a
few products gathered at selective retail outlets may not be representative of the full
array of goods sold. Also, many international surveys are undertaken to establish
differences in the cost of living experienced by business executives and their families.
These naturally focus on a set of products that are not representative of all purchases.
2.2 This Paper’s Method 6
Other studies have used price differentials as evidence of protection and to
estimate the benefits of integration. 7 This section discusses how I have tried to
overcome the challenges mentioned above, in order to produce improved estimates of
NTB food final goods protection and its effects. I use data in which every effort has
been made to ensure comprehensive coverage and comparability. In addition, I have
6
See Bradford and Lawrence 2004 and Bradford 2003 for more discussion of the methodology and data
presented in this paper and for welfare analyses of total protection.
7
See in particular Hufbauer et al 2002.
Page 8 of 39
endeavored to compare producer prices by eliminating the effects of distribution
margins. The data is also analyzed at a fairly disaggregated level, to mitigate weighting
problems.
I start with carefully matched retail prices that the OECD collects on a regular
basis in order to calculate purchasing power parity (PPP) estimates. With the
cooperation of member governments, OECD researchers build on the resources,
expertise, and data possessed by various national consumer price index (CPI) agencies
and sample prices of over 3000 final goods, about half of which are in food. They
make every effort to compare equivalent products across countries. For the most part,
they rely on identical brand names or exact descriptions of the items to be priced.
When they cannot find appropriate matches based on descriptions, researchers from
the nations involved travel abroad to determine which items would be most appropriate
matches for the items in their country. This has occurred with grain, some vegetables,
and tobacco. The researchers also call upon the expertise of producers, trade
associations, and buyers for large stores in order to determine matches.
Prices are collected from many markets and outlets at different times during the
year in order to obtain a single annual, national average (World Bank 1993, p10). Also,
prices of the average-sized purchase for that country are compared. After collecting
the data, apparent mismatches in quality are dealt with either by refining the
specifications or discarding the data (OECD 1995, p5). This method does not produce
perfect data, but the scale of resources expended on accurate matching indicates that
these are excellent measures of price differences for equivalent products.
Page 9 of 39
The researchers aggregate the most detailed price data into categories called
“basic headings”. These are defined as “groups of similar well-defined commodities for
which a sample of products can be selected that are both representative of their type
and of the purchases made in participating countries” (OECD 1995, p5). Thus, a basic
heading should not be too broad or too narrow. It should not be so broad that very
different products are compared; it should not be so narrow that few economies in the
sample sell it. For instance, seaweed is too narrow, and food is too broad.
In multilateral comparisons, one usually cannot find products that are
representative of the category and typical of what is bought in every country, since
consumers in different nations buy different mixes of products. Thus, while most items
are priced in most or all of the nations, not every product in the sample is priced in
each country. To be included in the sample, a product needs to be a “representative
product” in at least one country and it must be sold in large enough quantities in at
least one other country so as to be price-able. A “representative product” is one that
accounts for a large share of that country’s expenditure on that basic heading. For
instance, cheddar is a representative product for the cheese basic heading in France
but not for Italy. Cheddar cheese, however, is price-able in Italy. As long as nations
price their own major products and a share of all other products, relative prices for
each product and country can be calculated indirectly as well as directly. For details on
how the prices are combined into one average price for each country see EurostatOECD PPP Programme 1996. There are about 200 basic headings in the whole sample.
I obtained unpublished basic heading price data for 1999 and trimmed the sample to
Page 10 of 39
about 50 traded food goods. All prices were converted to US dollars using the 1999
exchange rates. See Table 1 for the list of categories.
The consumer price measures were converted to producer prices using data on
margins—wholesale trade, retail trade, transportation, and taxes—which come from
national input-output tables. 8 I did so for nine countries: Australia, Belgium, Canada,
Germany, Italy, Japan, the Netherlands, the United Kingdom (UK), and the United
States (US). Although I wanted to include more nations, such as France, the
availability of detailed margins data determined which ones became part of the sample.
I matched these margins with the OECD retail price data and derived estimates of
producer prices by peeling off the relevant margins. Thus,
[1]
p ijc
,
p =
1 + m ij
p
ij
pijp:
pijc:
mij:
the producer price of good i in country j,
the consumer price of good i in country j, as taken from the OECD data,
the margin for good i in country j, as taken from the national IO table.
Unfortunately, margins data only become available with a considerable time lag.9
The producer price estimates were therefore obtained by assuming that distribution
margins were the same percentage of overall value-added as they were in the most
recent year for which data were available.
8
Roningen and Yeats 1976 also use retail prices and adjust for taxes and transport costs, but they do not
adjust for wholesale and retail trade margins, which significantly outweigh taxes and transportation.
9
The margins data come from the following years: Australia, 1995; Belgium, 1990; Canada, 1990;
Germany, 1993; Italy, 1992; Japan, 1995; Netherlands, 1990; UK, 1990; and US,1992.
Page 11 of 39
Producer prices allow us to get a sense of which industries in which nations have
the lowest prices, but inferring the extent of insulation from foreign competition
requires one more step: taking account of transport costs from one nation’s market to
another. A foreign good must travel from the foreign factory to the foreign border and
then to the domestic border in order to compete with a domestic good. 10 Thus, one
cannot infer protection simply by comparing producer prices. The domestic producer
price must be compared to the import price of the foreign good. Such import price
data that matches the producer prices in this data does not exist independently and
needs to be inferred. This is done by combining data on export margins, also available
from national input-output tables, with international transport costs. 11
I could only get detailed data on international transport costs for Australia and
the US. Each reports import values for detailed commodities on both a basis that
includes insurance and freight (cif) and one that does not—so-called free on board
(fob). The cif/fob ratio is a good measure of all the costs of shipping goods from
abroad to these economies. The ratios for both nations are small, so that the gap
between the two is also small: the average for all products for the US is 1.05, while the
overall average for Australia it is 1.09. Thus, for each detailed sector, the average of
the two cif/fob ratios is used as an estimate for the international transport cost for that
product for all the countries.
10
For a discussion of the importance of export margins, see Rousslang and To 1993.
I have export margins for all countries except the UK, for which I used the Netherlands export
margins. Export margins tend not to vary much by country, so I feel confident that using the
Netherlands margins does not compromise the results.
11
Page 12 of 39
These data on export margins and international transport costs are used to
compute import prices for each product and country, as follows. Adding the export
margins to the producer prices enables one to calculate the export price for each
product in each country. The lowest export price plus the common international
transport cost is the import price. Thus, the export price is given by:
[2]
p ije = p ijp (1 + em ij ) ,
pije: the export price of good i for country j,
emij: the export margin of good i for country j.
The import price is then given by:
[3]
p iI = p iMe (1 + tri ) ,
piI :
tri:
p iMe
the import price of good i (the same for each nation),
the international transport margin for good i,
= min( p ie1 , p ie2 ,K , p ie9 ) : the minimum of the 9 export prices.
The ratio of each country’s producer price to the import price gives us an initial
measure of protection, pr ijIN :
[4]
prijIN =
p ijp
.
p iI
Page 13 of 39
For a given good, these measures will differ from true protection if all of the countries
in the sample have barriers to imports for that good. For such goods, the calculated
import price will exceed the true import price to the extent that the low cost producer
has barriers against imports. This will bias the protection estimates downward. By the
same token, if just one of the nine has no barriers to imports in that good, then prijIN
will approximate true protection, because, in this case, the price in the free trading
country will approximate the import price. Since the sample includes Australia, Canada,
and the US, which are fairly free traders, the low price in the sample approximates the
import price the great majority of the time.
Nevertheless, data on trade taxes are used to correct, at least partially, for the
possible downward bias. These tariff data come from the OECD tariff database, which
gives most favored nation tariff rates for member countries at the Harmonized 6-digit
level. The final measure of total protection, prijTOT , is given by:
[5]
prijTOT = max( pr ijIN ,1 + tar ij ) ,
tarij:
the tariff rate for good i in country j.
I simply use the fact that tariffs provide a lower bound on protection. If the initial
measures do not exceed the overall tariff rate, then that tariff rate is used as the
measure of protection. This happened about one-third of the time. After this
correction, the only time that these protection measures will be biased downward is
when all nations in the sample have NTBs against the rest of the world.
Page 14 of 39
These measures provide estimates of the protective effect of all kinds of
barriers—tariffs and NTBs alike. For our purposes, we want to focus on the impact of
NTBs alone, so I perform one final, simple modification. Tariffs are subtracted from
these total protection numbers. Mathematically, NTB protection is given by
[6]
pr ijNTB = pr ijTOT - tar ij = max( pr ijIN - tar ij ,1) .
Note that, since we measure protection as a ratio of the world price, a value of 1
indicates no protection. Thus, I conclude that there is no NTB protection whenever
pr ijIN - tar ij < 1 ⇒ pr ijIN < 1 + tar ij , that is, whenever the percentage by which the
producer price exceeds the import price does not exceed the tariff rate.
Figure 1 shows a schematic example that illustrates this methodology. Suppose
that there are three countries, with consumer prices as shown: Country A with the
lowest and Country C with the highest. C’s consumer price is nearly 2.5 times that of
A, but such a facile comparison can mislead. After peeling off domestic distribution
costs for this good, the ratio of C’s producer price to A’s is lower, though still large. As
is often the case in reality, in this example, the country with the high consumer price
also has the highest percentage domestic distribution margin. Converting to producer
prices gets us closer to our goal, since these provide a clearer indication of how
efficient producers in different nations are. Still, as discussed above, a straight
comparison of producer prices would overstate protection, since doing so would not
take account of the costs required to sell in foreign markets. So, to each of the
Page 15 of 39
producer prices, we add the unavoidable export margins and the international transport
costs. Note that, because of its relatively small export margin, Country B ends up with
the lower border price, even though its consumer and producer prices are higher than
A’s. In the end, the NTB protection level for C that we calculate is (25% - the tariff
rate) (if the tariff rate is lower than that), a much smaller gap than that between the
underlying consumer and producer prices.
3
SUMMARY AND ASSESSMENT
3.1 Four Key Characteristics
I believe that measures using this method, while not perfect, will shed useful
new light on NTB protection because they possess, to a large degree, four key
characteristics: completeness, comprehensiveness, accuracy, and international
comparability.
3.1.1 Completeness
Using price gaps enables one, in principle, to capture the combined effects of all
NTBs, which can include any number of regulations and bureaucratic procedures. For
example, a UN study analyzed how excess paperwork and cumbersome customs
procedures impede the international flow of goods. The study points out that, in
addition to direct costs, these regulations impose indirect costs, such as losses due to
“deterioration or pilferage” while cargo is waiting to be cleared, or the “strong
disincentive for potential exporters” imposed by complicated procedures. (See United
Page 16 of 39
Nations Conference on Trade and Development (1992).) The study estimated that
these barriers imposed costs that averaged 10 to 15%, on top of any other trade
barriers. Protection measures that rely on lists of individual barriers, such as the UN’s
own NTB measures, will tend to overlook subtle but real barriers such as these. This
paper’s method, however, will capture the protective impact of such barriers if they
raise domestic prices above the import price implied by the sample.
3.1.2 Comprehensiveness
These measures cover all traded final goods, instead of a small subset thereof.
Some other studies (such as Hufbauer and Elliott 1994) have limited their coverage to
sectors in which protection had been previously thought to exist, without testing
whether other sectors might enjoy well-disguised insulation from foreign competition.
The approach in this paper allows one to construct a more comprehensive picture of
final goods NTB protection in these nations. By the same token, this method does
exclude non-final goods, which account for most output and trade.
3.1.3 Accuracy
Accuracy stems from comparing actual prices of identical or equivalent goods.
Differences in quality have bedeviled attempts to use prices, except for certain
homogeneous goods. The data here, on the other hand, have resulted from intensive
multilateral efforts to correct for quality differences.
Page 17 of 39
3.1.4 International Comparability
Many other estimates have only been derived for a single country at a time,
making it difficult to rank economies in terms of openness. These measures use the
same data and apply the same method to each country in the sample, thus allowing
one to make such rankings, for individual products, for aggregated categories, and for
each country as a whole.
3.2 Possible Concerns
3.2.1 Imperfect Competition
Is it possible that market power could lead to estimates that do not really reflect
NTBs? I argue that this is not so. If the domestic producer price exceeds the
prevailing import price by more than the tariff rate, an NTB must support that gap, no
matter how those prices came to be. Market power does not change this fact. With
market power, a trade barrier may endogenously change prices, but the fact remains:
an un-arbitraged gap between the domestic price and the tariff-inclusive import price
cannot persist without NTBs that segment the domestic and world markets, and the
gap measures the amount of NTB protection.
3.2.2 Terms of Trade Effects
A related concern is the impact of terms of trade effects, for which this method
makes no adjustment. If an NTB drives down the import price, should one measure
Page 18 of 39
NTB protection with respect to the NTB-ridden import price or the free trade import
price? For instance, suppose that the latter is 1.00 and that a country imposes an NTB
of 0.2 that drives the domestic price to 1.10 and the import price to 0.90. Is the
amount of NTB protection 22% (
1.1
1.1
− 1 ) or 10% (
− 1 )? One could make an
0.9
1
argument for either, but this paper’s method presumes that the amount of NTB
protection is 22%, because that is the size of the wedge. With the barrier in place,
domestic consumers have to pay 22% more than people who can buy the good at
world prices. Consider a more extreme case. Suppose in the above example that the
domestic price remains at 1.00, while the import price gets driven to 0.80. It seems
that one should not conclude that NTB protection is zero simply because the domestic
price did not move; after all, the domestic price is 25% higher than the world price. In
practice, the terms of trade rarely, if ever, move as much as in the above examples and
will usually not matter. Even if one does want to correct for terms of trade effects, one
does not observe the free trade import price, so speculation would drive the correction,
and it would introduce a fair amount of uncertainty into the measures. Thus, for
theoretical and practical reasons, there is no correction for terms of trade effects.
3.2.3 Dumping
Dumping can possibly bias the inferred import price downward, which would bias
the protection measures upward. While protectionists make much of dumping, true
cases of dumping in which firms sell goods overseas below cost are rare to nonexistent. Most economists would agree that, the vast majority of the time,
Page 19 of 39
policymakers use anti-dumping duties as alternative ways to protect inefficient
industries, not as justified defenses against artificially low prices and the predatory
threats they pose. Even if such dumping occurs, and the resulting import price is lower
than otherwise, that does not invalidate it as a proper benchmark. Again, gaps
between domestic and import prices only result from barriers, even if the import prices
are artificially low.
3.2.4 Demand Differences
One may wonder whether these measures are valid if consumers in different
economies have different demands. The question arises: If Country A’s citizens have a
higher demand for good X than do Country B’s citizens, won’t that drive up the price of
good X in Country A in the absence of trade barriers? Answer: Only if there is a barrier
in Country A that allows such a gap to emerge. If Country A and Country B are truly
integrated, then good X will have one single demand curve, and the price will be the
same everywhere. Demand differences without barriers cannot sustain price gaps.
3.2.5 Price vs. Quantity Effects
Finally, in deriving these estimates, I realize that there is no clear connection
between tariff equivalents and the amount by which imports are reduced. Quantity
changes depend on market structure and such key parameters as the elasticities of
supply and of demand. Thus, a high NTB on a good with a low elasticity of demand
may reduce imports by less than a small NTB on a good with a high elasticity of
Page 20 of 39
demand. I do not purport, however, to analyze prices and quantities at the same time.
In order to assess the impact of the barriers on quantities, and thus on welfare, one
would need a model of the particular sector in question. I claim that the cleanest, most
effective way to measure NTB protection is to derive tariff equivalents and leave
quantity and welfare analysis for the next step.
4
THE EXTENT OF NTB PROTECTION
Table 2 presents the NTB data for the nine nations. Again, these are reported as
the ratio of the domestic producer price to the world price. Thus, a reading of 2.00
would be a protection rate of 100%. As mentioned above, the measures were
constructed using 50 categories, but, to facilitate the presentation, I have aggregated
up to 13 sectors, which correspond to the GTAP sectors that will be used in the AGE
analysis below. The table also reports weighted geometric means for each country. I
used the value of consumption as weights in constructing these means. Two factors
motivated this choice: 1) Protection skews the value of consumption less than
protection skews the value of production or of imports, and 2) The OECD reports the
value of consumption along with its price data, so consumption data that exactly
matches the protection aggregation was available. While these expenditure shares vary
by country, the first column of the table presents median expenditure shares to indicate
the importance of each sector.
These results imply that Canada, Italy, the Netherlands, and the US have the
lowest food NTB barriers, averaging less than 10%. Australia, Belgium, Germany, and
Page 21 of 39
the UK rank in the middle, ranging from 11% to 22%. Japan is a huge outlier: its food
NTBs average more than 90%, and Japan’s NTBs are the highest in each category
listed except garden products (mostly houseplants and planting products) and eggs.
Overall, this analysis suggests that there is nontrivial NTB food protection in industrial
nations, but Japan’s barriers loom very large. Much work needs to be done to bring
greater transparency and openness to Japan’s food markets.
Looking at individual sectors, these results imply that, in addition to Japan,
Germany, the UK, and the US have significant NTBs in fresh fruits and vegetables. As
one would expect, the data show Australia, Canada, and the US with low barriers in the
two meat sectors, while the Europeans and Japanese have extensive NTBs. In dairy
products, in addition to Japan, Australia, Belgium, Canada, and the US seem to have
nontrivial NTBs. Belgium and the UK have higher barriers than average in the large
processed food sector, but, once again, Japan’s barriers loom much larger than anyone
else’s. The beverages and tobacco sector probably has the most measurement error
because of the difficulties involved in correcting for large taxes. With this caveat in
mind, we have some evidence that Australia, Canada, and the UK join Japan with
significant barriers.
One may wonder about the sugar estimate for the US: 0% NTB protection.
Three factors contribute to this result. First, Australia is the low-price producer sample,
but its import price is probably higher than the true world price, biasing sugar
protection estimates downward. Second, to make its sugar restrictions more WTOcompatible, the US has converted its quotas to tariff-rate quotas, which means that its
Page 22 of 39
official tariff rate is high (about 75%; see Table 3 below). The tariff rate ends up
exceeding the inferred NTB price gap, resulting in a finding of no NTB protection.
Finally, the underlying price data only includes sugar sold to final demand, not sugar
sold to food processing firms; the price gaps for final demand sugar are probably lower
than the gaps for sugar sold to produers.
For comparison purposes, Table 3 provides tariff data. Not surprisingly, tariffs
are generally lower and more tightly distributed. The Europeans have the highest
average tariffs in food. One can use the tariff and NTB numbers to calculate a measure
of “protection transparency”, which is defined as the ratio of tariff protection to total
protection (which is simply the sum of NTB and tariff protection). These data imply
that Japan and Australia have the most opaque food protection regimes, while Italy and
the Netherlands have the most transparent.
Obstfeld and Rogoff 2000 concludes that “a recurring theme here is that the
markets for most ‘traded’ goods are not fully integrated, and segmentation due to
various trade costs can be quite pervasive. In fact, the spectrum of goods subject to
low trade costs may be very narrow.” Our data provide support for this view in the
realm of food.
5
POLICIES BEHIND THE PRICE GAPS
These NTB estimates may help policy makers in one of two ways. First, for
known NTBs, these measures provide estimates of the extent to which those NTBs
actually restrict trade. Thus, these results may provide useful information to trade
Page 23 of 39
negotiators as they decide how to efficiently focus their efforts on freeing up trade.
Second, some sectors that have not reached the trade negotiation agenda may, in fact,
enjoy significant disguised NTB protection that is worth negotiating down. This
research can help to flag such sectors.
To illustrate how these results can help in the first way mentioned, Table 4
shows possible barriers for some of the NTB gaps, though much more work along these
lines needs to be done. I have drawn on the EU Market Access Database, the USTR’s
2000 Report on Foreign Trade Barriers, and 2000 WTO Trade Policy Review for the
European Union, the US, and Japan. A more detailed analysis would reveal more
policies behind the NTBs. Also, for any given price gap, the policies listed may not be
major causes, but they are initial candidates.
Looking back at Table 2, there are a number of NTBs for which there are no
listed possible policies. In these cases, more detailed research may reveal particular
sources of the gaps, which might then become subject to negotiation. Also, any of
these gaps, as well the ones which have listed policies, could result from burdensome
customs procedures and other administrative friction, as discussed above. Thus, efforts
by trade negotiators to remove such widespread sand from the wheels of trade could
potentially have large benefits across many sectors and economies.
6
THE WELFARE EFFECTS OF INTEGRATION
To provide insights into the importance of NTBs, this section simulates their
removal. For eight of the nine nations, the simulations compare real incomes in the
Page 24 of 39
world as it is with one in which the NTBs are eliminated. (Unfortunately, data problems
prevent Belgium from being included.) I use an AGE model based on one developed by
Harrison, Rutherford, and Tarr (HRT). 12 The model has considerable country and
sectoral detail: 16 regions and 33 sectors (See Table 5). 13 The model also allows for
both increasing returns to scale and dynamic adjustment of the capital stock. The next
two subsections describe the model and then report the simulation results.
6.1 Description of the Model
6.1.1 Production Structure
Production involves the use of intermediate goods and five factors—capital,
skilled labor, unskilled labor, land, and natural resources. Only capital can move across
national boundaries; all factors can move freely across sectors. Value added in each
sector has a CES (constant elasticity of substitution) production function. This
formulation means that, within each sector, the elasticity of substitution between any
two of the factors is the same. I use HRT’s values for these elasticities, which they
estimated econometrically using US time series data from 1947 to 1982 and using the
same functional form as is used in this AGE model. In their estimates, however, they
used only three factors—capital, labor, and land—instead of five. See Table 6 for these
12
The model is based on the computer code provided by Glenn Harrison, Thomas F. Rutherford, and
David Tarr. Their code is available for public access at http://theweb.badm.sc.edu/glenn/ur_pub.htm
and was used in their 1995, 1996, and 1997 articles.
13
The underlying data come from Version 5 (1997) of the Global Trade Analysis Project (GTAP)
database.
Page 25 of 39
estimates and their standard errors. The production function for intermediates and the
value-added composite is Leontief. 14
Some sectors are assumed to have constant returns to scale. Other sectors,
though, are modeled with increasing returns to scale and imperfect competition. 15 In
these sectors, there is firm-level product differentiation, with output being a composite
of varieties. Firms have fixed costs and constant marginal costs, meaning that reducing
the number of firms leads to rationalization gains. These firms compete using quantity
conjectures, with entry and exit that drive profits to zero.
Dynamics are incorporated by allowing the capital stock to vary in response to
changes in the rate of return caused by liberalization. If the rate of return increases,
investment increases the capital stock until its return is driven back down to the longrun equilibrium. The results, therefore, reflect the model’s predictions for what
happens after the capital stock has changed enough to return the price of capital to its
original level. The capital adjustment process is not modeled, and the time horizon
implied by these results depends on how long one thinks it takes capital to respond to
interest rate differentials. The model ignores the consumption foregone by the
increased investment, which may overstate the estimated benefits. On the other hand,
the model ignores any impact of growth on productivity and innovation, which leads to
an underestimate of the gains.
6.1.2 Demand Structure
14
Relaxing this assumption does not significantly change the results.
Page 26 of 39
On the demand side, each region has a representative consumer and a single
government agent, each of whom has a nested CES utility function and practices multistage budgeting. At the top level, demand across the 33 sectors is Cobb-Douglas.
Consumers first decide how much to spend on each of the 33 aggregate goods, given
total income and aggregate prices. Each of these goods is a CES composite of
domestic output and an import composite, which are imperfect substitutes. In this
second level, consumers divide spending between the domestic and import good by
maximizing a CES utility function subject to the total spending they have allocated to
that sector and given the aggregate prices in that sector. At the third level, the model
invokes the Armington assumption in that imports of the same good from different
economies are assumed to be imperfect substitutes. Preferences across these different
goods from different economies are given by a CES utility function. At this third level,
consumers choose quantities of each import subject to the amount they have budgeted
for aggregate imports at the second level and subject to the various prices. I follow
HRT and set the elasticity of substitution across import varieties, σ MM , equal to eight
and the elasticity of substitution between the import composite and the domestic good,
σ DM , equal to four. These elasticities affect the magnitude of the results. Higher
values of these parameters lead to greater substitution in response to price reductions
and, in general, higher welfare gains from liberalization. Roughly speaking, cutting
these elasticities in half reduces the gains by 10% to 50%, depending on the region
and the simulation. Similarly, doubling these elasticities increases the estimated gains
15
See Table 6 for the sectors and the mark-ups used. This table also presents alternative mark-ups from
Page 27 of 39
by about 20% to 100%. Even such wide changes in the calibration, however, do not
change any of the main conclusions.
In the sectors with increasing returns, yet another level of constrained choice is
introduced. In this set-up, the domestic good and each import good produced in each
region, instead of being homogeneous goods, are themselves composites of different
varieties produced by the different firms. Consumers have CES preferences over these
varieties and allocate spending across them subject to the amount they budgeted for
each good at the third level. The elasticity of substitution across these varieties is set
at 15. All results are robust to wide changes in this parameter.
6.1.3 Incorporating the New Data
6.1.3.1
Protection Data
To simulate the impact of NTBs, the model was benchmarked with the total
protection measures—NTBs plus tariffs—instead of the GTAP protection data, which
consists almost entirely of tariffs. In the model, all policy distortions enter as ad
valorem price wedges 16 , which, conveniently, is the form that this paper’s new
protection data take. So, replacing the GTAP tariff equivalents with these data is fairly
straightforward. I did not, however, simply use the new measures as is, since they
apply only to final goods, while all of the sectors of the model contain a combination of
final and intermediate goods. Instead, I used a weighted average of the new data and
the original GTAP data. The weight on the former was the fraction of output in that
the GTAP model. The results are robust to the set of mark-ups used.
Page 28 of 39
sector sold to final demand; the GTAP measure got the complementary weight. Thus,
letting B and GTAP be the two protection measures and α , the final demand fraction,
the protection estimate used was αB + (1 − α )GTAP . Using this method ensures that
model sectors with a high proportion of final goods use a protection estimate close to
mine, while sectors with a low fraction of final goods use a protection estimate close to
the GTAP measure. Put another way, the lower the final demand fraction, the less the
data deviated from the standard GTAP data. See Table 7 for a comparison of these
weighted data and the original GTAP data. This table shows the food estimates in bold;
it also includes total protection estimates for non-food sectors in the model. Replacing
the GTAP data with these does not significantly affect the food results.
6.1.3.2
Distribution Margins Data
The margins data used to derive the protection measures allow one to model
distribution more accurately within the AGE framework. Most AGE trade models do not
account for margins explicitly. All distribution services are lumped into the trade and
transport sector and consumed as a separate good, instead of being linked to the
goods that use those distribution services. Since margins vary across sectors, this
obscures the role of distribution in the economy and can skew the results of AGE
analyses. For instance, simulations of price reductions in other sectors may imply a
large substitution out of trade and transport services, even though actual consumption
of these will probably increase in order to facilitate commodity flows. Also, not
16
Government revenue is held constant throughout all simulations by assuming that lump-sum taxes are
used to replace any lost tax revenue.
Page 29 of 39
accounting for margins implies that consumers base choices on producer prices instead
of the higher consumer prices that include margins.
These problems are addressed by incorporating distribution explicitly into each
final demand sector for which there is margins data. This is done by treating margins
like taxes, since margins create a wedge between consumer and producer prices. For
the eight nations involved, therefore, margin wedges were insertedinto each of the
relevant sectors. 17 The value of the trade and transport sector was reduced by the
total value of these margins. Finally, inputs into the trade and transport sector were
reduced and re-distributed across the final goods sectors in accordance with the
amount of distribution used in those sectors. 18
6.2 Welfare Analysis
This section presents estimates of the potential gains from including food NTBs
on the trade negotiation agenda. Since tariffs presumably require much less work to
remove, it is not likely that negotiators will remove NTBs and not tariffs. So two sets of
scenarios are simulated: one in which nations remove all food protection—NTBs and
tariffs alike—and one in which nations only remove tariffs. For each of these two
situations, I conduct two types of simulations: unilateral barrier removal in each of the
eight nations and multilateral worldwide opening by all eight at once. I focus on
17
See Gohin 1998 and Komen and Peerlings 1996 for other examples of modeling margins in this way
within AGE models. Bradford and Gohin 2006 explicitly model the distribution sector for the US within an
AGE model.
18
These modifications only apply to final goods. Due to lack of data, I do not modify the model to
account for intermediate distribution. It turns out that these intermediate margins are quite a bit smaller
than the margins for final goods.
Page 30 of 39
changes in equivalent variation (which, given the model structure, is the same as
changes in real consumption) as a percentage of GDP.
Tables 8 and 9 show the main results for total protection and just tariffs. These
tables report the permanent, annual effect of trade opening on consumption, as a
percentage of GDP, once the capital stock has changed to its new equilibrium.
Alternatively, they report the welfare costs, born at home and abroad, of tariff and total
protection in the eight nations separately and as a group. Table 10 shows the
difference between the two scenarios and thus the predicted extra gains from removing
food NTBs. (Alternatively, the results in Tables 8A and 8B are simply the sum of the
results from Tables 9 and 10.) For each table, Panel A reports these gains as a
percentage of GDP, while Panel B shows them in billions of 1997 US dollars.
Tables 8A and 8B imply that, overall, food protection in these eight nations
imposes significant costs on the world. If all food barriers in all eight were removed,
world welfare would increase by 0.73% of GDP, or about $185 billion (in 1997 dollars).
About $135 billion of that would accrue to rich countries, with less developed countries
(LDCs) getting permanent annual gains of about $50 billion. All but Canada, Germany,
and the US would reap significant gains from unilateral opening, and all but Germany
would benefit greatly from opening in all eight. Germany suffers from adverse terms of
trade effects when all eight open: the reallocation of resources causes demand for the
goods that Germany tends to export to decline, relative to demand for the goods that
Germany tends to import. Japan’s food barriers impose large costs on poor countries.
Every poor region would benefit most from Japanese opening. (“Rest of Europe” is
Page 31 of 39
mostly rich countries.) Interestingly, the US would get significant gains from Japanese
food barrier removal but not from its own.
Tables 9A and 9B reveal that poor countries would reap most of their gains from
the removal of tariffs, not NTBs: about $33 billion from tariff removal, compared to $50
billion for all protection removal. Food tariffs in Japan, Germany, the UK, and the US
impose the largest burdens on poor countries. Since tariffs are much easier to reduce
than NTBs, it appears that poor countries will get more bang from their negotiating
buck by focusing on food tariff removal in rich countries, rather than food NTBs.
Focusing on Table 10, most of these nations do not get significant extra annual
boosts to GDP from unilateral food NTB opening. Multilateral opening from all eight,
however, would bring nontrivial annual gains of at least 0.1% of GDP for all except
Canada and Germany. Global GDP would rise an additional 0.4%, or $90 billion, with
NTB removal in just the food sectors. It appears that poor countries have little to gain
from NTB reductions in Europe or the US. Japan, however, remains a large burden,
and their food NTBs warrant close attention from poor and rich countries alike. In fact,
Japanese opening of food NTBs accounts for almost as many gains as having all eight
open.
Three main forces drive the gains for any given country: the amount of
protection removed, the share of trade in GDP for that country, and terms of trade
effects. The US’s relatively low barriers and its low trade/GDP share lead to relatively
low predicted gains for the US. Similarly, the Netherlands’ high trade share amplifies its
percentage gains. On the other hand, Japan’s NTBs are so high that it reaps
Page 32 of 39
substantial extra gains from NTB liberalization despite the fact that Japan has the
lowest trade share in the sample: only about 10%. Terms of trade changes mute gains
for Canada, Germany, and Italy.
For all economies except Canada and Japan, the extra gains from multilateral
food NTB opening are significantly more than the gains from unilateral opening. These
six economies have incentives to engage in multilateral NTB reform, as opposed to
going it alone.
Overall, these results imply that the potential gains to be reaped from food
protection are not trivial, whether one considers tariffs or NTBs. Of course, such
extensive liberalization in these nations is not on the table right now. Complete
opening may not be an option because of short run political stresses caused by
contraction in protected sectors. Our analysis does not provide a recipe for reform, but
it does show that the potential gains from future attempts to integrate markets remain
quite large.
These estimates of the benefits of integration do not take account of certain
costs. In particular, differences in national languages, policies, and institutions may
well create barriers to price arbitrage, but they may also provide benefits that would be
lost if the world economy was to be deeply integrated in the sense we are exploring in
this study. Also, I have not considered adjustments costs as workers and other factors
move out of shrinking sectors into expanding sectors.
On the other hand, these results may understate the costs of the barriers by
treating them as if they were tariffs. There are at least two ways in which the costs of
Page 33 of 39
NTBs may be higher. First, they do not generate revenue for the government as tariffs
do, so this benefit is foregone. An NTB such as excessive fumigation raises costs to the
foreign exporter without necessarily generating income for the importing country. The
fumigators in that country may get paid more than otherwise, but this is a transfer from
within the economy, not an extra source of revenues as with tariffs. Second, removing
barriers may actually save resources and therefore yield even larger benefits than
estimated here. As Anderson and van Wincoop 2002 emphasizes, tariffs generate
deadweight losses, but NTBs may consume resources directly. Suppose, for example,
that two nations each require meat to be certified as safe even though their criteria are
very similar. Firms that wish to sell in both markets must expend real resources to
meet foreign requirements. Meat approved in one economy cannot simply be sold
abroad. Under these circumstances, in addition to the gains from removing the
barriers, freeing the resources that are consumed by the (unnecessary) duplicative
regulatory processes could produce additional gains.
The estimates are also conservative because they ignore the potential benefits
from opening nations outside the sample of eight used in the study.
7
CONCLUSION
This paper has presented a method for estimating food NTB protection in rich
countries. The estimates imply that rich nations harbor significant food NTB protection,
in addition to food tariffs. Japan has unusually high food NTBs. AGE simulations imply
that negotiating the removal of food NTBs, especially in Japan, would bring large
Page 34 of 39
benefits to rich and poor nations alike, implying that the extra work required to open
these markets would probably pay off.
Of course, the trade opening devil lurks in the details, so trade analysts need to
determine the actual policies that underlie the protection we have quantified in this
paper. It is easy for governments to claim that certain policies in other economies act
as trade barriers; the more difficult task is to provide evidence for these claims. We
have taken an initial step toward this goal by matching up suspected policies with
sectors for which we have evidence of NTB protection. As shown in Table 4, we find
that, for agriculture and food products, overly restrictive phytosanitary and sanitary
requirements, apparently unfounded import bans of certain products, and onerous
labeling rules emerge as potentially damaging trade barriers and worthwhile targets of
negotiations. Various experts for individual sectors are probably well aware of such
barriers; this paper has provided potentially valuable information by putting numbers
on the extent and effects of a wide range of barriers.
This initial analysis could be improved in a number of ways. More recent price
data are available, making it possible to derive updated protection data. Including
more countries in the price comparisons would improve the accuracy of the barrier
estimates and would provide a more complete picture of the potential gains from trade
opening. The AGE analysis could be improved by accounting for technological change.
Also, it would probably be worthwhile to add confidence intervals to the AGE estimates,
something which is quite feasible.
Page 35 of 39
I hope that this paper has provided useful initial insights on the extent of, the
effects of, and the policies underlying food NTB protection in rich countries. I also
hope that this paper will stimulate much-needed future research in this area.
Page 36 of 39
REFERENCES
J.E. Anderson and E. van Wincoop. 2002. “Borders, Trade and Welfare” in S.M.
Collins and D. Rodrik, eds. Brookings Trade Forum 2001. Washington DC:
Brookings Institution.
S.C. Bradford. February 2003. "Paying the Price: Final Good Protection in OECD
Countries." Review of Economics and Statistics. 85(1):24-37.
S.C. Bradford and A. Gohin. 2006. “Modeling Distribution Services and Assessing
Their Welfare Effects in a General Equilibrium Framework.” Review of
Development Economics. 10(1):87-102.
S.C. Bradford and R.Z. Lawrence. February 2004. Has Globalization Gone Far
Enough? Washington, DC: Institute for International Economics.
Eurostat-OECD PPP Programme. 1996. "The Calculation and Aggregation of
Parities." Unpublished.
J.A. Frankel. 1997. Regional Trading Blocs in the World Economic System.
Washington DC, Institute for International Economics.
A. Gohin. 1998. Modelisation du Complexe Agro-alimentaire Francais dans un Cadre
D’equilibre General. PhD Dissertation, Universite de Paris I, Panthéon-Sorbonne.
G.W. Harrison, T.F. Rutherford, and D. Tarr. December 1995. "Quantifying the
Outcome of the Uruguay Round." Finance and Development. 32(4):38-41.
G.W. Harrison, T.F. Rutherford, and D. Tarr. 1996. "Quantifying the Uruguay
Round" in W. Martin and L.A. Winters, eds. The Uruguay Round and the
Developing Countries. New York: Cambridge University Press.
G.W. Harrison, T.F. Rutherford, and D. Tarr. September 1997. "Quantifying the
Uruguay Round." Economic Journal. 107:1405-1430.
G.C. Hufbauer et al. 2002. The Benefits of Price Convergence: Speculative
Calculations. Washington, DC: Institute for International Economics.
G.C. Hufbauer and K.A. Elliot. 1994. Measuring the Costs of Protection in the
United States. Washington, DC: Institute for International Economics.
M.M. Knetter and P.K. Goldberg. 1995. “Measuring the Intensity of Competition in
Export Markets.” Cambridge, MA: National Bureau of Economic Research
Working Paper.
M.H.C. Komen and J.H.M. Peerlings. 1996. “WAGEM: an Applied General
Equilibrium Model for Agricultural and Environmental Policy Analysis.”
Wageningen Economic Papers: 1996-4, The Netherlands
M. Obstfeld and K. Rogoff. 2000. “The Six Major Puzzles in International
Macroeconomics: Is There A Common Cause?” NBER Macroeconomics Annual
2000. Cambridge, MA: NBER.
Organization for Economic Cooperation and Development (OECD). 1995.
Purchasing Power Parities and Real Expenditure. Paris: OECD.
V. Roningen and A. Yeats. 1976. "Nontariff Distortions of International Trade:
Some Preliminary Empirical Evidence." Weltwirtschaftliches Archiv. 112(4):613625.
Page 37 of 39
D.J. Rousslang and T. To. 1993. "Domestic Trade and Transportation Costs as
Barriers to International Trade." Canadian Journal of Economics. 26(1):208221.
United Nations Conference on Trade and Development (UNCTAD). 1992.
Analytical Report by the UNCTAD Secretariat to the Conference. New York:
United Nations.
World Bank. 1993. Purchasing Power of Currencies: Comparing National Incomes
Using ICP Data. Washington, DC: World Bank.
Page 38 of 39
FIGURE 1
NTB PROTECTION CALCULATION: SCHEMATIC EXAMPLE
CONSUPRODMER
Domestic UCER
Export
PRICE
Margin
PRICE
Margin
c
p
( pij )
( m ij )
( pij )
( em ij )
p =
p
ij
COUNTRY A
COUNTRY B
$2.24
$2.70
60%
80%
EXPORT
PRICE
( pije )
pijc
International LANDED
Margin
PRICE
( tm i )
( p il )
pije = pijp (1 + emij )
1 + mij
10%
50%
$1.40
$1.50
$2.10
33%
$2.31
10%
$2.00
$2.20
e
COUNTRY C
Î
$5.50
NTB Protection in C =
100%
Minimum Export Price: piM
e
pil = piM
(1 + tm i )
$2.75
p iCp
2.75
− tar ij =
− tar ij = 1.25 − tar ij , ie, (25% - tariff rate), if the tariff rate is
l
2.20
pi
25% or less. Otherwise, NTB protection is inferred to be zero.
Note: i indexes products, and j indexes countries.
Page 39 of 39